**5. Discussion**

As the problems of the greenhouse effect and global warming are becoming increasingly serious, reducing carbon emissions has become an international consensus. Thus, in this paper, we proposed suggestions for low-carbon optimization of the land-use spatial distribution pattern based on an analysis of the effect of carbon emissions from land-use changes and the prediction of future tendencies in Jinhua. Many studies have explored the characteristics of land-use carbon emissions through coefficients of carbon emissions of land use in different regions and analyzed the existing problems in the land-use pattern [26–29]. Compared with previous studies, this study gives more practical advice on carbon emission reductions and proposed an optimized land-use spatial pattern using a multivariate linear programming model and a FLUS model with local driving factors.

In this study, we selected carbon emission coefficients in accordance with the actual situation in the study area. For instance, the carbon emission coefficient of woodland was selected with consideration of the carbon sink of the main forest vegetation types and the carbon emissions produced through firewood collection, forest fires, and HWP emissions of forest products in Zhejiang Province. This is more accurate than using the average coefficient at the national scale. However, the temporal effects of carbon emissions from changes in the age of stands, the site index, and the stand structure were still neglected. In terms of calculation of the carbon flow, we just considered natural and manmade carbon emissions but did not consider changes in carbon storage in land ecosystems. This may have caused the results to deviate from the actual situation. In addition, land use was divided into six types based on the first LUCC classification system, but the differences in carbon emissions intensity among different land-use types under the secondary classification were not considered. Thus, the spatial distribution of carbon emissions was not accurately described. Additionally, without driving factors, we were only able to perform pure mathematical simulations through the CA–Markov and FLUS models. The results do not reflect the influences of social and economic factors on the evolution of the land-use pattern well. Thus, in this study, driving factors, such as population, economy, transportation, and terrain were added to the models, and the impact of planning policies in the study area was also considered. The optimized simulation result in 2030 in Jinhua shows that in order to maintain stable economic performance under the premise of strictly controlling the intensity of land development, we should not only aim to limit the rate of construction land growth in terms of quantity but also to promote the construction of the Jinhua-Yiwu New Area and Jinhua-Lanxi Innovation City guided by the development of axial belts and a group layout. This will give full play to the connectivity of land development axes and promote the compound use of land and the intensive use of resources from the perspective of the spatial layout.

In future research, the carbon emission coefficient should be modified to improve the accuracy of the accounting results. Additionally, the impacts of land-use change on both carbon emissions and carbon storage should be considered. Our prediction models could be improved by refining the land-use classification so that the method can be applied to small-scale research more accurately in the future.
